Physical Database Structure{5}

by David H
The article that I read this week is about building a best fit data warehouse. This article talks about the structure of data warehouse and why physical database design is crucial. Data warehouse structure has four crucial models. There are symmetric multi-processing (SMP), cluster, massively parallel processing (MPP), and grid. The goal of these models is to process a large amount of data. As an architect, they need to come up with solution on how to measure disk space, processing power, and network bandwidth. In business, the requirement of database express on Business Intelligence workload. The dimension of time and data volume is the most challenging of structure. Basically, what data warehouse does is to serve the users who were performing a lot of tasks.  During that time, all of transaction database need to send operational report in real time. The physical design of data warehouse focuses on three terms. There are Operational Business Intelligence, Enterprise Business Intelligence, and Historical Business Intelligence. The characteristic of Business Intelligence workloads include data volume, number of users, type of query, frequency and timing of access, and latency requirement. These characteristics are very important of physical design that influence in data warehouse.

This article relates to class because it mentions on how the physical designs of database affect to data warehouse. I have learned that the characteristic of Business Intelligence workload is very important to database structure. However, it was little complicated but as long as we follow step by step of physical design, we will be fine.  I think we should spend more time on physical database design topic. We need to have deep understanding about this in order to build database.



O’Brien, J. (2008). Building a best-fit data warehouse: Why understanding physical database structures matters. Business Intelligence Journal, 13(1), 51-62.